c_prog/main.c

#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>

const u_int NAME_MAX_LENGTH = 32;
const u_int TAG_MAX_LENGTH = 32;
const u_int HASH_BUCKETS = 4000037;

int string_compare(const char *str1, const char *str2) {
    while (*str1 && *str2 && *str1 == *str2) {
        str1++;
        str2++;
    }
    return *str1 - *str2;
}

void string_ncopy(char *dest, const char *src, size_t max_len) {
    size_t i = 0;
    while (i < max_len - 1 && src[i]) {
        dest[i] = src[i];
        i++;
    }
    dest[i] = '\0';
}

typedef struct person {
    struct person *next;
    char name[32];
    int count;
} person;

void newPerson(person *p, const char *name) {
    string_ncopy(p->name, name, NAME_MAX_LENGTH);
    p->count = 1;
    p->next = NULL;
}

void insert(person **head, person *node) {
    if (head == NULL) {
        *head = node;
    } else {
        person *p = *head;
        int p_exists = string_compare(p->name, node->name);
        while (p->next != NULL && p_exists != 0) {
            p = p->next;
            p_exists = string_compare(p->name, node->name);
        }
        if (p_exists == 0) {
            p->count++;
        } else {
            p->next = node;
        }
    }
}

void sorted_name_insert(person **head, char *name) {
    person *node = (person *) malloc(sizeof(person));
    newPerson(node, name);
    if (*head == NULL) {
        *head = node;
    } else {
        person *p = *head;
        person *p_prev = NULL;
        int p_exists = string_compare(p->name, name);
        if (p_exists > 0) {
            node->next = *head;
            *head = node;
            return;
        }
        while (p->next != NULL && p_exists != 0) {
            p_exists = string_compare(p->next->name, name);
            if (p_exists > 0) {
                node->next = p->next;
                p->next = node;
                return;
            }
            p = p->next;
        }
        if (p_exists == 0) {
            p->count++;
            free(node);
        } else {
            p->next = node;
        }
    }
}

void display(person *head) {
    person *p = head;
    while (p != NULL) {
        printf("%s %d\n", p->name, p->count);
        p = p->next;
    }
}

u_long hash(const unsigned char *str) {
    u_long hash = 5381;
    int c;
    while ((c = *str++)) {
        hash = ((hash << 5) + hash) + c;
    }
    return hash;
}

void hm_insert(person **hashmap, char *name) {
    u_long hash_value = hash(name);
    hash_value = hash_value % HASH_BUCKETS;
    sorted_name_insert(&hashmap[hash_value], name);
}

char *parse_line(char *line) {
    char *line_it = line;
    if (*line_it == '<') {
        line_it++;
        char *tagname = malloc(sizeof(char) * TAG_MAX_LENGTH);
        char *tag_it = tagname;
        while (*line_it != ' ' && *line_it != '>' && *line_it != '\0' && *line_it != '\n') {
            *tag_it = *line_it;
            line_it++;
            tag_it++;
        }
        *tag_it = '\0';
        if (string_compare(tagname, "author") == 0 || string_compare(tagname, "editor") == 0) {
            free(tagname);
            while (*line_it != '>') {
                line_it++;
            }
            line_it++;
            char *last_space = line_it;
            while (*line_it != '<') {
                if (*line_it == ' ') {
                    last_space = line_it;
                }
                line_it++;
                if (isdigit(*line_it)) {
                    line_it = line_it - 2;
                    while (*line_it != ' ' && *line_it != '<') {
                        line_it--;
                    }
                    last_space = line_it + 1;
                    break;
                }
            }
            char *name = malloc(sizeof(char) * NAME_MAX_LENGTH);
            char *name_it = name;
            line_it = last_space + 1;
            while (*line_it != '<' && *line_it != ' ') {
                *name_it = *line_it;
                line_it++;
                name_it++;
            }
            *name_it = '\0';
            return name;
        }
        free(tagname);
    }
    return nullptr;
}

void clean_memory(person **hashmap) {
    for (int i = 0; i < HASH_BUCKETS; i++) {
        person *p = hashmap[i];
        while (p != nullptr) {
            person *next = p->next;
            free(p);
            p = next;
        }
    }
    free(hashmap);
}

int main(void) {
    const auto hashmap = (person **) malloc(sizeof(person *) * HASH_BUCKETS);
    for (int i = 0; i < HASH_BUCKETS; i++) {
        hashmap[i] = NULL;
    }
    FILE *fp = fopen("dblp.xml", "r");
    char line[1024];
    while (fgets(line, 1024, fp) != NULL) {
        char *name = parse_line(line);
        if (name != NULL) {
            hm_insert(hashmap, name);
            free(name);
        }
    }
    fclose(fp);
    for (int i = 0; i < HASH_BUCKETS; i++) {
        display(hashmap[i]);
    }
    clean_memory(hashmap);
    return 0;
}